[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: FW: sming meeting minutes draft... Questions for the wg.

At 09:00 AM 12/18/2002 -0600, sidney.antommarchi@tekelec.com wrote:

>>Further, attendance at the SMIng/EOS meetings in Atlanta was close to
>>I attribute this to lack of interest and my guess is that this is because
>of XML.
>Although I share you enthusiasm for XML as a technology for management, the
>reality is that I have not seen any RFP related to NGN networks in the last
>3 years that mentions XML at all.
>Attendance and/or support from companies for standards work is directly
>related to the state of the economy.  Justifying travel has been very
>difficult this year for everyone.  Other standards bodies are having the
>same problem.  As always, network management takes second place to call
>processing and budgets reflect that.
>>I agree, but I think vendors would update their products if the cost was
>>enough and the improvement in data transfer efficiency was high enough.
>I have not done this in a while but the last time that I checked,
>industrial grade XML tool sets where prohibitively expensive compared to
>SNMP ones.  But, this is not the only driver here.  Network operators like
>their HP OV NNM, Micromuse, etc.  To change the existing toolsets is going
>to take some serious selling.  Why should the operators spend a dine on
>XML?  In this age of huge pipe  over capacity, I do not think that data
>transfer efficiency would be a good selling point.  Specially, considering
>that network management traffic is so minuscule as to be unimportant in
>traffic studies.

I agree data transfer efficiency is not a primary factor,
and hence, the verbosity of XML encoding is not a major concern,
for the applications that XML can provide better solutions than SNMP.

>Yes, it is possible to convert SMI to XML. So?  What is the problem that is
>being solved?  I hope that y'all are not doing this for the simple reason
>than because it can be done.  Can anyone state the problem?

Who said the goal was to convert SMIv2 to XML?  What a waste of time.
XML data structures are 100 times more powerful than SMIv2.
Consider the possibility that XML is not related to SNMP at all,
but rather targeted at CLI applications.  Most network operators
today are using CLI for configuration management, not any of the
SNMP applications you mention above.  I have no doubt that SNMP
will continue to be used for monitoring applications for several
more years. 

There have been several I-Ds written about why operators aren't 
using SNMP for configuration, so I won't rehash that argument here.  
Many operators are even using CLI for monitoring, and syslog for 
event-driven fault monitoring. XML can be used to provide a
better programmatic interface than CLI in all 3 scenarios.

>Sidney Antommarchi
>VXi - Systems Engineering


>                      Andy Bierman                                                                                                      
>                      <abierman@cisco.         To:      "Glenn Waters" <gww@nortelnetworks.com>                                         
>                      com>                     cc:      eos@ops.ietf.org                                                                
>                      Sent by:                 Subject: Re: FW: sming meeting minutes draft... Questions for the wg.                    
>                      owner-eos@ops.ie                                                                                                  
>                      tf.org                                                                                                            
>                      12/18/2002 12:15                                                                                                  
>                      AM                                                                                                                
>At 10:43 AM 12/17/2002 -0500, Glenn Waters wrote:
>>I meant to send this to the list...
>>-----Original Message-----
>>From: Waters, Glenn [CAR:IO47:EXCH]
>>Sent: Monday, December 16, 2002 23:30
>>To: 'Durham, David'
>>Subject: RE: sming meeting minutes draft... Questions for the wg.
>>Where to start.
>>XML is overshadowing all other network management discussions in the IETF.
>It is hard to do anything SNMP without XML being mentioned. Further,
>attendance at the SMIng/EOS meetings in Atlanta was close to pathetic. I
>attribute this to lack of interest and my guess is that this is because of
>I think there is very little official IETF activity related to XML.
>There is a lot of XML interest and activity by the vendors and their
>customers, especially compared to interest in SNMP.
>>SMIv3 provides much needed structure to the data; however, I fear the
>forklift upgrade necessary in order to get it deployed will be very
>detrimental to its success. I am not convinced SMIv3 could be deployed
>before XML has traction. I also don't think that using SMIv3 with other
>protocols such as XML is useful -- there are too many other schema
>definition languages for XML already. Further, if there are simplifications
>that we need for the other XML schema languages we should pick one and work
>with that body.
>This is a valid criticism of SMIv3.  It would impact everybody, especially
>the readers and writers of MIBs.  Moving to XML would also be a major
>change for readers and writers.  I don't know what you mean by a choice
>of XML schema languages.  I think the choice has to be XML Schema from
>the W3C.  The only other choice is DTDs, and they are not expressive enough
>to foster interoperability or allow for complex, reusable data types.
>>I think that the OOPS draft has the most potential although we should
>ensure the features that it offers are what we want to pursue. OOPS does
>not require a forklift upgrade. The features that it offers can operate on
>the existing SMI. The one issue with EOS is that an upgrade to the protocol
>stack is still required. Given the amount of time this is taking for SNMPv3
>I fear that industry uptake will be below an acceptable level.
>I agree, but I think vendors would update their products if the cost was
>enough and the improvement in data transfer efficiency was high enough.
>SNMP is mostly used for monitoring, and it is horribly inefficient at
>transferring data, especially for a binary encoding.
>>The bottom line is that interest in SMIng and EOS is low. Interest in XML
>is high. SNMP has no workers. XML has people jumping to help. I think that
>the IETF needs to reconcile where we want to go in the network management
>space. The reality is that XML is coming, it just may be that the IETF will
>not be defining it.
>I agree, but I hope the IETF will be clueful enough to pay attention
>to this trend.  (BTW, I've noticed that the people that dismiss XML
>and all the tools work around it as just another fad are the least
>likely to actually have read the XML specs or checked out the tools.)
>>> -----Original Message-----
>>> From: Durham, David [<mailto:david.durham@intel.com>mailto:david.durham@intel.com]
>>> Sent: Monday, December 16, 2002 16:53
>>> To: 'Wijnen, Bert (Bert)'; sming@ops.ietf.org
>>> Subject: RE: sming meeting minutes draft... Questions for the wg.
>>> Originally there was much excitement around the smi-ds proposal from the
>>> wg
>>> participants. Now that we are peeling back the onion a bit I think we
>>> seeing some of the complexities of this approach emerge, and we are
>>> experiencing some notable feature creep as well. So I wonder if the wg
>>> sentiment is cooling and why. Two fundamental questions that I would
>>> people to (quickly) respond:
>>> Are people still interested in perusing the hierarchical instance naming
>>> using oids as proposed in the smi-ds document?
>>> Are you interested in helping by being an editor on one or more of the
>>> documents listed below?
>>> If feature creep and gratuitous changes are the only issues, that is
>>> fixable. I want to understand if the fundamentals of the smi-ds proposal
>>> are
>>> still viable and have wg support.
>>> -Dave
>>> > -----Original Message-----
>>> > From: Wijnen, Bert (Bert) [<mailto:bwijnen@lucent.com>mailto:bwijnen@lucent.com]
>>> > Sent: Monday, December 16, 2002 4:48 AM
>>> > To: Durham, David; sming@ops.ietf.org
>>> > Subject: RE: sming meeting minutes draft...
>>> >
>>> > I'd like to get a few more details on this (I think we got them
>>> > at the meeting):
>>> >
>>> > > Revisited Charter and Milestones. Updated charter was put on
>>> > > the mailing list. No issues raised on the list, no issues
>>> > > raised at the meeting either. According to the charter
>>> > > milestones we are a year behind. The original milestones
>>> > > assumed the nmrg documents which were complete, but the wg
>>> > > chose to investigate the smi-ds route. We will still need a
>>> > > few iterations on the smi-ds/v3 documents before they will be
>>> > > complete. The current proposed list of documents in priority order
>>> > >
>>> > > -        1. SMIv3 Language Definition: Andy Bierman
>>> > > -        2. Capabilities MIB: Andy Bierman: DONE
>>> > > -        3. SMIv3 Guidelines
>>> > > -        4. Transition from SMIv2
>>> > > -        5. SMIv3 MIB Modules (core types)
>>> > > -        6. INET Modules (textual conventions)
>>> > > -        7. RFC 2580 Conformance Updates
>>> > >
>>> > > We need volunteers for the documents or the wg will shut
>>> > > down, and the smi will not progress. Previous volunteers for
>>> > > the guidelines and transition documents are waiting for the
>>> > > language definition. In principle, people support the smi-ds
>>> > > work, but we will need people to sign up to get the work
>>> > > done. Likewise from the discussion at the wg meeting it seems
>>> > > that there is a lot of waffling on how we proceed item by
>>> > > item through the open issues listed below.
>>> > >
>>> > My understanding at the meeitng was that Andy un-volunteered for some
>>> > docs. And so I like to clearly understand who is currently volunteer
>>> > for what. I also like to see commitments as follows:
>>> >
>>> > - Volunteers to commit to deliver a reasonable revision of the
>>> >   document they volunteer for
>>> > - From the WG to do serious review and to provide serious input
>>> >   for the editors and the chair, so that we can get to consensus.
>>> >
>>> > The lack of WG participation and the lack of enthusiastic volunteers
>>> > for all documents does not bode well. And as David said, we're
>>> > already 1 year behind schedule.
>>> >
>>> > Thanks, Bert